WO2008101752A1 - Method for determining the fatigue of a pump rotor of a gas turbopump - Google Patents

Abstract

A method for determining the fatigue of a pump rotor of a gas turbopump has the following method steps: continuous detection of the rotor rotational speed (n) of the pump rotor, detection of the local rotational-speed maxima and minima of a temporal rotational-speed profile under consideration, association of the rotational-speed maxima and minima with one another to form pairs, detection of a pair fatigue value (L) for each of the rotational-speed pairs, and accumulation of all pair fatigue values (L) to form an overall fatigue value (Ltot). It becomes possible in this way to detect the cycle loading for the pump rotor of a vacuum pump and to incorporate it into the calculation of an overall fatigue value.

Description

 Method for determining the fatigue of a pump rotor of a turbo gas pump

The invention relates to a method for determining the fatigue of a pump rotor of a high-speed turbo gas pump,

Turbo gas pumps and in particular turbo-molecular pumps are operated at relatively high speeds of 10,000 - 100,000 rpm. The pump rotor of the gas pump is subject by the high centrifugal forces, temperature influences, etc. of a mechanical fatigue. Especially in turbo-gas pumps narrow gaps are provided between the pump rotor and the pump stator, to get the highest possible output. The pump rotor fatigue causes a failure of the pump rotor or a collision between the pump stator and the pump rotor after a certain time. To avoid this, the pump rotor is replaced before the estimated collision time.

It can be provided a time constant maintenance intervals, which provides for maintenance or replacement of the pump rotor after a certain time or a certain period of operation.

From DE 101 51 682 Al a method is known in which from the absolute speed and the temperature of the pump rotor, a dynamic maintenance interval! is calculated. As a result, the so-called creep stress of the pump rotor is detected and taken into account in the calculation of the dynamic maintenance interval. Investigations have shown that in this way, under certain operating conditions, a practical dynamic maintenance system! can not be determined with sufficient accuracy.

The object of the invention is therefore to provide a method for determining the fatigue of the pump rotor, which determines the fatigue of the pump rotor as close to the reactor as possible for all operating conditions actually occurring.

This object is achieved with the features specified in claim 1,

According to the method according to the invention for determining the fatigue of the pump rotor of a turbo-gas pump, the following method steps are provided:

Continuous determination of the rotor speed (n) of the pump rotor, Determination of the local speed maxima and minima of a considered speed curve,

Assigning the speed maxima and minima to each other in pairs,

• determining a pair fatigue value (L) for each of the speed pairs, and

• Accumulate all pair fatigue values (L) to a total fatigue value (! «, [■).

Experiments have shown that independently of a possible creep stress the so-called Zyciusbeanspruchung significantly into the life or in the fatigue of a scheldldrehenden metal component with enters. Only when the cycle load is additionally detected can, in particular, the risk of fractures on the pump rotor in the determination of the mechanical fatigue of the pump rotor realistically,

Cycle stress already occurs under stresses well below the so-called tensile strength R _{m of} a metallic material, and often below the yield strength R _βl which characterizes the onset of macroscopic plastic deformation. The reason for the cycle stress lies in the interaction of the microstructure of the respective material with the cyclic stress load. Also below the yield strength microscopic microplastic deformations can occur in the material. These can be done on the one hand at local stress concentrations in the component, but can also occur in softer structural areas, for example in a precipitation-hardened material. However, microscopic deformations can also occur in the region of individual crystals, which are oriented crystallographically favorably to the direction of stress, so that they are locally localized to one another - A -

Slip line and thus can lead to plastic deformation. As these effluents emerge from the component surface, microscopic step formation occurs, also called intrusions and extrusions. These represent notches, which lead to a local voltage overshoot. At this point, then in the course of further cyclic stresses, a crack occurs, which is driven so long by the cyclic stresses until the component is so weak at this point that a component break occurs.

To detect and estimate the cycle stress, it is necessary to observe the rotor speed of the pump rotor and to detect the end or completion of a speed cycle. In this case, a speed cycle is to be understood as meaning a time interval at the beginning and at the end of which the sign of the acceleration of the pump rotor has changed, that is to say the time course of the pump rotor acceleration has a zero crossing. A speed cycle preferably lasts from an IMull passage of the temporal Beschleunigungsverfaufes to the following Nuildurchgang. A speed cycle may alternatively be understood to mean a time interval beginning with the beginning of a positive acceleration and ending with the end of a subsequent negative acceleration.

In order to determine the so-called Zyciusbeanspruchung of the pump rotor, the rotor speed (π) of the pump rotor is continuously determined In this way, a time-speed Veriauf created, for which the respective local speed maxima and speed minima are determined. A speed maximum or minimum is in each case a speed extreme value, after which the sign of the acceleration changes. After reaching a maximum speed, so the speed decreases, so after reaching a speed minimum, so increases the speed. The considered period may refer to a single cycle, which may last from a maximum to a minimum or from a minimum to a maximum from the start to the stop of the pump rotor can refer to a fixed period or to a set number of maxima and minima.

For the considered period of the speed curve, the speed maxima and minima are assigned to each other in pairs. Different rules or algorithms can be used for the assignment of the speed maxima and minima to each other for the generation of maximum-minimum pairs.

From each speed maximum-minimum pair a pair fatigue value L is determined. Finally, all determined pair fatigue values L are accumulated to a total _{fatigue value} L _tσt .

For a complex speed curve, therefore, first after a specified time! determines each speed pair, calculates a fatigue value for each speed pair, and accumulates the calculated fatigue values to a total fatigue value. Thus, the total fatigue value determined in this way indicates the total fatigue due to the cycle stress of the pump rotor in the considered period. In this way, the possibility is created to calculate a maintenance interval in the calculation of the cycle load of the pump rotor is included. A maintenance interval can thus be determined with a high degree of safety and considerably more accurate, so that fatigue-related damage on the one hand and large safety discounts for determining the maintenance interval on the other hand can be avoided.

Preferably, a characteristic mechanical stress of the respective speed pair is included in the determination of the wear value L. The characteristic voltage of the speed pair can be, for example, the arithmetic mean of the two mechanical stresses caused by the Centrifugal forces occur in the considered component at the two speeds of the maximum speed-Minimurn pair. In this way, the speed level and the speed increase are included in the determination of the wear value.

Preferably, the rotor speed is determined in constant time intervals. The continuous determination of the rotor rotational speed in the present case therefore also means a determination in constant discrete time intervals. A time interval may be one or more seconds. This is usually sufficient, since the speed of, for example, turbo gas pumps is not constantly changing and in particular not very snowy! can be changed. Introducing a reasonable time interval to determine the rotor speed reduces the load on the computer hardware involved.

According to a preferred embodiment of the invention, the pair-fatigue value is determined by an at least quadratic, more preferably a cubic polynomial. With this, a realistic fatigue value can be calculated, which indicates the incurred stress sufficiently accurately. Alternatively, however, the pair fatigue value may also be more easily determined if some increased inaccuracy is accepted.

Preferably, a fatigue signal is output as soon as the total fatigue value exceeds a total fatigue limit. For example, the alarm fatigue limit can be selected so that there is still a reserve left until the end of the maintenance interval so that maintenance can be prepared with a lead time of several days or weeks. Of course, multiple total fatigue limits may be provided which cause various warnings towards the end of a maintenance interval.

In the following an embodiment of the invention will be explained in more detail with reference to the drawings. Show it :

1 is a diagram showing the relationship between the Spanπungsamplitude σ _π corresponding to a respective speed and the number of cycles to destruction (Wöhler Dϊagramm),

Fig. 2 is a schematic representation of the relationship between

Speed cycle and the voltage cycle with the assignment of σ _a and other specific voltage characteristics to the speed, and

Fig. 3 shows a possible cyclic operation e.g. a turbo gas pump, shown as speed n over time f to define various parameters.

When operating a turbo-gas pump and in particular when operating a turbomolecular pump high radial mechanical stresses in the form of voltages occur due to the high speeds and the high centrifugal forces, especially in the pump rotor, given the smallest possible gap between the pump stator and the pump rotor It can come even with relatively small mechanical changes of the pump rotor to a collision between the pump rotor and the pump stator.

In addition to the influence of the pure speed and the rotor temperature, which is described in DE 101 51 682 Al and describes the so-called creep, also the so-called cycle stress goes into the mechanical fatigue of the pump rotor. Experiments have shown that the material fatigue by creep on the one hand and by so-called cycle stress on the other hand in a usual practice stress range of stress and temperature is largely independent of each other. The temporal succession of the creep stress and the cycle stress is not of great importance when considering the total fatigue of a component, so that both stresses can be added together after their determination.

In Fig. 1, the relationship between the voltage amplitude σ. represented on the ordinate and the number of corresponding speed cycles to destruction on the abscissa. From this it is clear that there is a limit for σ _π , below which the number of sustainable cycles tends towards infinity until it breaks. The relationship between speed and voltage is given by material, shape, mass, etc. of the pump rotor, so that both sizes can be expected alternatively. In the following, the mechanical stress is used as an example to determine the fatigue.

In FIG. 2, for better understanding of the following algorithms, an example of a cycle progression as a function of the pump rotor speed n and of the corresponding voltage σ over time t is shown.

For each cycle variant of a turbo-gas pump, a curve corresponding to FIG. 1 or an entire set of curves for different stresses is determined experimentally. From these curves, the constant factors of a polynomial can be determined, with which it can be determined for each cycle variant according to FIG. 2 how many uniform cycles would occur up to the destruction in total:

\ og (N _lot ) = α "+ A - σ _m + β ₂ - σ _m ² + β, ^■ σ) _n + γ _r σ _B + χ ₂ ^• σl + ri - σl With

N _ιπt = number of uniform cycles (voltage _swings ) to destruction

_σ = - i L = mean stress

σ _a = σ _h - σ _m = σ _m ~ σ _t = voltage swing / 2 = amplitude voltage

zugeordneten Drehzahl-Wertecr _lt high voltage; n _ZΛ and n _{z <2} are the other

assigned speed values

f minui. , " ,)]. .. _. σ, = - σ _ref . _{ykl! is} - - - = undervoltage; σ _ref is the tension in the rotor

at the speed n _rcf

The coefficients ß _x ~ ß _% and γ _λ - γ ₃ and a ₀ are determined by interpolation, for example by the method of the sum of the squares of errors, from experimentally determined value triplet N _tot , σ, _{B /} σ _a .

As can be seen, the average stresses σ _m and the stress stroke σ _π - 2 are included in the calculation via the mechanical stresses.

FIG. 3 shows by way of example a temporal speed curve from practice. The continuous determination of the rotor speed n takes place in constant time intervals Δt. Of course, At can be considerably shorter fail, as shown here, to realize a more closely meshed speed determination and to obtain a more accurate picture of the temporal speed curve. After falling below a minimum speed or Mϊndestdrehfrequenz, for example, 10 Hz, a fatigue determination is made.

As soon as the sign of the acceleration changes, the speed maximum or minimum that has just been passed is stored as the local maximum or minimum. If the stored speed is a local maximum, it is stored in a maximum memory. If the stored speed is a local minimum, it is stored in a minimum memory. In the present example, in this way at the points B ₁ D, F, H maxima and at the points A, C ₁ E, G Minima found.

In order to detect the cyclic load, temporally adjacent maxima-minima pairs are not formed, but extreme value pairs are formed over the observation period. The first pair consists of the smallest speed minimum and the highest speed maximum, in this case of the pair AD, the second pair consists of the remaining extreme values G ₁ F, the third pair of the values C and B etc .. This approach has established itself as a proven method.

Speed values, which represent a minimum in the time sequence and would have to be used as the maximum due to the value pair formation, thus do not constitute a true maximum and fall out of the overall rating. The same applies accordingly for false minimums. In the present case this is the case for the pair E, H.

The period of time considered for the fatigue determination may be a complete cycle of operation, that is from the last time the gas pump is switched on until the gas pump is switched off. The considered period can but also be the entire pump life. The considered period will in practice be limited by the available computing and storage capacity. Considering the total pump life so far, the total accumulated wear value is recalculated every time.

For each maximum-minimum speed pair, the paired fatigue value L is calculated using the formula resolved after N _tot according to the formula:

L = VN _M

The total fatigue value L _tot then results from the sum of the pair fatigue values L:

L ₁₀₁ = Z lI N ₁₀₁₁ = ^

For this purpose, furthermore, the accumulated creep fatigue value can be continuously added. The accumulated cycle fatigue value L _m or the combined creep cycle fatigue value can be continuously compared with a total fatigue limit value L _max . The limit value L _mgx is less than 1.0, since a threshold value L _max = l, 0 a failure is expected. As soon as the total fatigue value L _tot exceeds the limit value L _max , a corresponding fatigue signal is output.

Claims

claims 1. A method for determining the fatigue of a Purnpenrotors a turbo gas pump, comprising the steps of: Continuous determination of the rotor speed (n) of the pump rotor, Determination of the local speed maxima and minima of a considered speed curve, Assigning the speed maxima and minima to each other in pairs, • determining a pair fatigue value (L) for each of the speed pairs, and Accumulating all pair fatigue values (L) to a total fatigue value (L _tot ), 2. A method for determining the fatigue of a pump rotor of a turbo gas pump according to claim 1, characterized in that at least one characteristic voltage of the speed pair is included in the determination of the pair fatigue value (L). 3. A method for determining the fatigue of a pump rotor of a turbo-gas pump according to claim 2, characterized in that a characteristic voltage is the average voltage of the respective speed-pair. 4. A method for determining the fatigue of a pump rotor of a turbo-gas pump according to any one of claims 1-3, characterized in that a local speed maximum or minimum is a Nuiidurchgang the time course of the rotor acceleration 5. A method for determining the fatigue of a pump rotor of a turbo-gas pump according to any one of claims 1-4, characterized in that the determination of the rotor speed [n] in constant time intervals (.DELTA.t) takes place. 6. A method for determining the fatigue of a pump rotor of a turbo gas pump according to any one of claims 1-5, characterized in that the pair fatigue value (L) is determined by an at least quadratic, more preferably at least cubic polynomial. 7. A method for determining the fatigue of a pump rotor of a turbo gas pump according to any one of claims 1-6, characterized in that a fatigue signal is output as soon as the total fatigue value (L _tot ) exceeds a Gesarntermüdungs threshold (l _ma χ). 8. A method for determining the fatigue of a pump rotor of a turbo gas pump according to any one of claims 1-7, characterized in that continuously creep fatigue values from the rotor speed (π) and the rotor temperature determined and accumulated in the total fatigue value (L _tot ).